Ask anyone at Shift Technology what makes them proud to work for the company, and they’ll likely say it’s the positive impact their work has on customers and the global insurance industry every day. The company’s AI-powered decision optimisation solutions and fraud detection tools help the world’s leading insurers mitigate fraud and risk throughout the policy lifecycle, streamline claims processing, improve operational efficiency, and deliver exceptional customer experiences.
To make this possible and sustainable, AI and machine learning have been at the heart of Shift Technology since its inception, beginning with its first use case: claims fraud detection. Insurance companies often face costly and complex challenges in detecting fraud—especially as fraudsters continue to evolve and outmanoeuvre existing systems. While technology progresses to detect fraud more effectively, so too does the technology enabling fraudsters to inflict even greater harm—often amounting to billions of pounds each year.
Shift Technology is committed to staying at the forefront of AI and machine learning, continually exploring innovative ways to support its clients—insurance companies and others—facing constantly evolving fraud schemes and operational challenges. Because data is central to everything it does, the company’s data scientists work closely with customers to understand their workflows and help them get the most from the platform. Insurers are then empowered to make smart, informed decisions—backed by Shift Technology’s expert analysis.
To make onboarding easier for clients, Shift Technology standardised its platform on Microsoft Azure. The result is a robust, efficient cloud infrastructure that allows its engineering and data science teams to fully leverage new capabilities—regardless of location. “Given the global footprint of Azure and its broad capabilities, the core of the Shift platform runs on a Microsoft technology stack that allows us to operate and innovate simultaneously,” says Marc Jones, Chief Technology Officer at Shift Technology.
One key innovation is the company’s use of Azure OpenAI Service, part of Azure AI Foundry, along with Azure AI infrastructure, to refine its generative AI capabilities and help make insurance decisions faster, more accurate, and more transparent. “Because Microsoft is at the forefront of generative AI adoption and we’re on Azure, we can move rapidly from research to real proof of concept and straight into production using Azure OpenAI,” Jones adds.
With every solution—including newly launched tools for claims automation, underwriting risk mitigation, and subrogation recovery—Shift Technology processes and generates large volumes of documents in varied formats, languages, and structures. To date, the company has analysed more than 2.6 billion policies and claims, along with their supporting documentation. Historically, staff spent weeks labelling data and training models to generate useful outputs. “A significant amount of unstructured data enters our platform as documents, so being able to automate classification and information extraction with high accuracy is essential,” says Jones.
Using machine learning–based optical character recognition (OCR) with cloud-based Azure AI Vision—also part of Azure AI Foundry—and both OCR and layout detection via Azure AI Document Intelligence, Shift Technology streamlines document classification and rapidly extracts relevant information. “By reducing weeks of work to days, we’re now able to fully utilise Azure OpenAI and focus on the prompt engineering that enables data extraction,” continues Jones. “We can then fine-tune prompts to get outputs that previously took much longer to achieve.” This has enabled the team to build a reusable prompt library and continuously adapt it to new document types.
Given the scale of its services and data volume, Shift Technology uses Azure Kubernetes Service (AKS) for containerised deployments to help scale parts of the platform. “We still run core compute-heavy workloads on VMs or containers in AKS,” explains Jones. “But when it comes to adopting more cloud-native services, orchestrations or workflows, we’re keen to leverage Azure Service Bus for eventing, Azure Container Apps for simple container deployments, and Azure Functions for focused workloads.”
In this new environment, Shift Technology can quickly deploy new AI endpoints via Azure OpenAI for immediate use. The team also values Azure OpenAI’s extensive and regularly updated model catalogue. They use GPT-series models in production and benchmark them against others in Azure AI Foundry, which streamlines testing and comparison. Serverless deployment of language models is especially helpful for testing capabilities securely and cost-effectively, without needing to provision large or expensive VMs.
By the nature of their industry, Shift Technology’s insurance clients require data hosting within specific regions to ensure security and compliance. To meet these demands, the company needed its AI and machine learning capabilities to run natively in designated Azure regions. Historically, Shift ran its AI models on infrastructure it managed itself, allowing full visibility and control over data flows, including data residency and usage. Although newer generative AI tools offer options for self-hosting and training, they are often complex—particularly for compute- and GPU-intensive applications.
“We wanted to take advantage of the latest generative AI tools, ideally with a hosted solution, but we also had to meet stringent security and compliance requirements,” says Jones. “Bringing Azure OpenAI into our Azure ecosystem gave us the ability to have meaningful discussions with Microsoft about how data is handled, secured, and the implications for compliance.”
Likewise, Shift Technology can clearly communicate with its clients, assuring them that their data is not used for training models—an important reason behind choosing Azure OpenAI. The company also leverages global Azure regions and enhances resiliency through multiple availability zones. “We’re seeing a snowball effect—more teams and more use cases—because people recognise the impact of these technologies,” says Jones. “It’s exciting to see teams embrace innovation to enhance our services for customers.”
With success across internal teams and insurance clients alike, Shift Technology plans to expand its use of generative AI to enhance existing capabilities and introduce new ones, including current exploration into agentic systems. It has already begun using GitHub Copilot and is assessing the potential for increased developer efficiency and productivity. “For us, it’s less about writing code and more about augmenting developers’ ability to do their jobs and simplifying the tasks around coding,” says Jones. “Above all, I want our continuous innovation to position us as a leading AI company—always using and exploring the latest technologies to serve our clients better.”